91ɬÂþ

Abstract

In recent years, photoplethysmography (PPG) is considered extensively because it is a low-cost electro-optical technique for measuring the volumetric variations of blood and contains valuable health information such as heart rate (HR), respiration rate (RR), blood pressure (BP), and blood oxygen saturation (SpO2). However, PPG sensing in wearable devices remains constrained by two primary challenges for long-term deployment: (1) the high-power consumption of the light-emitting diodes (LEDs) and (2) the requirement for high dynamic range in signal acquisition. In order to enable energy-efficient, wearable, and long-lasting health monitoring devices this work addresses these challenges by proposing two innovations: (1) a No-LED (ambient light) mode for PPG signal acquisition to significantly reduce power consumption, and (2) a current-mode analog-front end (AFE) architecture designed to achieve ultra-low-power operation while maintaining high dynamic range performance.

In the first part of the work a flexible wireless finger patch capable of recording PPG signals under diverse ambient lighting conditions (No-LED) is researched. Experimental evaluation on 12 subjects demonstrated that under sufficient ambient light such as sunlight, room light, the acquired PPG signals can be placed in the reliable or acceptable category and can measure HR and BP with good accuracy. The calculated maximum mean absolute error (MAE) for HR in comparison to the reference measurement is small (3.4 BPM). For BP prediction an average MAE of 8.1 mmHg and 6.05 mmHg for systolic BP (SBP) and diastolic BP (DBP), respectively are obtained.

As a user can be without enough ambient light for a significant portion of the day, it is important to have PPG sensor-based wearables that work in both No-LED and LED-ON mode light. Moreover, the reported finger patch cannot measure SpO2, which is an important health parameter. To overcome these limitations, in the second part of the work a ring–bracelet flexible wearable device is explored. The wearable integrates an optimized photodiode–LED configuration to work in both No-LED (when there is enough ambient light) and low intensity LED modes (when there is not enough ambient light) and an accelerometer for motion artifact reduction. It also utlizes optical filters for SpO2 measurement. This system achieved a maximum measurement error of 3 bpm for HR and 2.9% for SpO₂, while continuously consuming only 262.5 μW in No-LED mode and 652.5 μW in low-LED mode).

While the wearable devices explored in first and second part demonstrate the feasibility of No-LED operation, their readout circuits were not optimized for weak PPG signals under ambient light or low-LED illumination. To address this limitation, a novel current-based readout architecture (I-readout) is developed. The designed PPG signal readout works entirely in the current domain, eliminating the need for high-performance closed-loop amplifiers and enabling integration with other current/voltage-output sensors. The architecture achieves an 86 dB dynamic range while consuming only 70 μW in the AFE. Total system power remains as low as 150 μW with the low-LED active and 70 μW in No-LED mode, representing the first application of current-mode signal by minimizing or eliminating active LED usage and leveraging efficient current-domain signal processing, this work collectively lays the foundation for a new generation of ultra-low-power, scalable, and sustainable wearable health monitoring technologies by PPG sensing.